The major objective of this investigation is to conduct a multi-center, double-blinded randomized trial comparing the efficacy of two therapies for the prevention of recurrent spontaneous abortion (RSA): immunotherapy using paternal leukocytes combined with "Tender Loving Care" (TLC) therapy versus TLC therapy alone. We propose to randomize 524 couples over four years. Couples meeting the following criteria will be included in this trial: l)greater than or equal to consecutive spontaneous abortions; 2) No more than one viable gestation (greater than or equal to 28 weeks gestation) in the first pregnancy only; 3) No maternal lymphocytotoxic antibodies (LCTA) or maternal serum blocking factors (BF); 4) No known cause of miscarriage (excluding cytogenetic, autoimmune, anatomic, infectious etiologies); 5) No contraindications for using husband's cells for immunization; and 6) No consanguinity between partners or increased risk for a malformed fetus. Couples meeting the study criteria will be randomly assigned into one of two treatment groups and be immunized with either husband's cells or saline. All couples will receive TLC therapy weekly during gestational weeks 4-12, consisting of blood drawing, physical exam, ultrasound exam, and 15-20 minutes with the patient's physician. A success is considered to be achieving pregnancy within 12 months of randomization/immunization and delivering a viable offspring (greater than or equal to 28 weeks gestation). A failure is considered to be not achieving pregnancy within 12 months of randomization/immunization or experiencing a fetal loss after pregnancy is achieved. Serum samples will be stored before and weekly during the first trimester for studies of LCTA and BF. All couples achieving pregnancy and their neonate or abortus will be typed for HLA region genes using DNA probes. Comparisons between the two treatment groups will be made with respect to the following variables: 1) success rates; 2) pregnancy rates; 3) production of LCTA or BF; and 4) the presence of birth defects or chromosomal abnormalities in liveborn and abortuses, respectively. Associations between pregnancy outcome and maternal-fetal HLA status, and production of LCTA and BF will also be assessed. All analyses will consist of binomial comparisons of proportions in the two groups, corrected for continuity, or by logistic regression to assess simultaneous effects of several variables.